open-llm-leaderboard-old/details_GeorgiaTechResearchInstitute__galpaca-30b
收藏数据集概述
数据集来源
该数据集是在评估模型 GeorgiaTechResearchInstitute/galpaca-30b 在 Open LLM Leaderboard 上的自动创建的。
数据集组成
数据集由64个配置组成,每个配置对应一个评估任务。数据集从4次运行中创建,每次运行可以在每个配置中找到特定的分割,分割名称使用运行的时间戳。"train" 分割始终指向最新的结果。
额外配置
一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示在 Open LLM Leaderboard 上的聚合指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_GeorgiaTechResearchInstitute__galpaca-30b", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-15T01:05:49.975308 运行的最新结果: python { "all": { "em": 0.205746644295302, "em_stderr": 0.00413985910625819, "f1": 0.288852768456377, "f1_stderr": 0.004183984738478157, "acc": 0.3265751062277813, "acc_stderr": 0.009076887028812184 }, "harness|drop|3": { "em": 0.205746644295302, "em_stderr": 0.00413985910625819, "f1": 0.288852768456377, "f1_stderr": 0.004183984738478157 }, "harness|gsm8k|5": { "acc": 0.028051554207733132, "acc_stderr": 0.00454822953383636 }, "harness|winogrande|5": { "acc": 0.6250986582478295, "acc_stderr": 0.013605544523788008 } }
配置详情
-
config_name: harness_arc_challenge_25
- split: 2023_07_19T22_01_48.453969
- path: **/details_harness|arc:challenge|25_2023-07-19T22:01:48.453969.parquet
- split: 2023_08_09T12_18_52.169485
- path: **/details_harness|arc:challenge|25_2023-08-09T12:18:52.169485.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2023-08-09T12:18:52.169485.parquet
- split: 2023_07_19T22_01_48.453969
-
config_name: harness_drop_3
- split: 2023_09_23T10_10_35.995390
- path: **/details_harness|drop|3_2023-09-23T10-10-35.995390.parquet
- split: 2023_10_15T01_05_49.975308
- path: **/details_harness|drop|3_2023-10-15T01-05-49.975308.parquet
- split: latest
- path: **/details_harness|drop|3_2023-10-15T01-05-49.975308.parquet
- split: 2023_09_23T10_10_35.995390
-
config_name: harness_gsm8k_5
- split: 2023_09_23T10_10_35.995390
- path: **/details_harness|gsm8k|5_2023-09-23T10-10-35.995390.parquet
- split: 2023_10_15T01_05_49.975308
- path: **/details_harness|gsm8k|5_2023-10-15T01-05-49.975308.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2023-10-15T01-05-49.975308.parquet
- split: 2023_09_23T10_10_35.995390
-
config_name: harness_hellaswag_10
- split: 2023_07_19T22_01_48.453969
- path: **/details_harness|hellaswag|10_2023-07-19T22:01:48.453969.parquet
- split: 2023_08_09T12_18_52.169485
- path: **/details_harness|hellaswag|10_2023-08-09T12:18:52.169485.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2023-08-09T12:18:52.169485.parquet
- split: 2023_07_19T22_01_48.453969
-
config_name: harness_hendrycksTest_5
- split: 2023_07_19T22_01_48.453969
- path:
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-human_aging|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-international_law|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-management|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-marketing|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T22:01:48.453969.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_
- path:
- split: 2023_07_19T22_01_48.453969



